Re: target/5610: Not clear how to get FP operations to use SSE

Re: target/5610: Not clear how to get FP operations to use SSE

[hubicka]

Synopsis: Not clear how to get FP operations to use SSE

State-Changed-From-To: open->closed
State-Changed-By: hubicka
State-Changed-When: Thu Oct 10 10:54:17 2002
State-Changed-Why:
    Updated for 3.2.1 and 3.3

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=5610

target/5610: Not clear how to get FP operations to use SSE

[tom]

>Number:         5610
>Category:       target
>Synopsis:       Not clear how to get FP operations to use SSE
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          doc-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Feb 06 06:26:00 PST 2002
>Closed-Date:
>Last-Modified:
>Originator:     Thomas Womack
>Release:        gcc-3.1-20020204
>Organization:
>Environment:

>Description:
The option "-msse-math" appears in doc/invoke.texi (line 484, in the list of i386 and x86-64 options), but the compiler does not recognise it.

This option is now spelt "-mfpmath=sse".

I think SSE2 should be mentioned explicitly as an instruction-set extension in line 7766 (along with MMX, SSE and 3dNow), and (though -mfp-math=sse does reference -msse2) that the documentation for -msse should include a line like "to have SSE/SSE2 instructions generated automatically from floating-point code, see @option{-mfpmath=sse}". Also, -msse2 should be added to the @opindex entries around that line.
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: